Assessing Consistency of Scenarios Across Scales Developing globally linked internally consistent scenarios under the Shared Socioeconomic Pathways framework

Abstract

In global environmental change research, anticipating the implications of large-scale environmental changes on local development is an important endeavour for mitigating and adapting to difficult challenges. Researchers have used multi-scale scenario analysis to anticipate future changes. Simply put, multi-scale scenario analysis is used to model cross influences between factors or drivers operating at different scales, for example, global, regional, and national levels. To ensure that scenarios are plausible, which is important for policy decisions, scenarios must be consistent across scales. However, there is confusion to what cross-scale consistency means. Consistent scenarios across scales refers to how lower level (e.g., national) scenarios should be developed considering various development pathways at the global scale that can potentially influence domestic developments. Scenario studies often use the term ‘consistent’ as defined by Zurek and Henrichs’ (2007) linking strategies. Zurek and Henrichs (2007) categorize different strategies for linking scenarios across scales. The categorization is based on the process by which scenarios developed by different modelling teams are linked. The degree to which these scenarios are linked is characterized as equivalent, consistent, coherent, comparable, and complimentary—with equivalent as the strongest link, whereas complimentary as a weak or no link. Link strength is defined by how similar (or different) the scenario elements (logics, drivers, assumptions) are. Linking scenarios across scales (e.g., global and regional) should aim to be equivalent or consistent across scales; this can be achieved by quantitative downscaling. For scenarios developed in parallel, the degree to which these scenarios can be viewed as consistent depends on whether the elements in these scenarios are the same, if not similar. However, adhering to this criterion is challenging because lower level scenarios may require different scenario elements to be incorporated in the scenario development process—these elements are factors or drivers that are operating at a more localized scale. Therefore, constraining the selection of scenario elements for developing regional or national level scenarios may be impractical. There are varying degrees of consistency of scenarios across scales much like the concept proposed by Zurek and Henrichs (2007) that spans from equivalent to complimentary. However, there is a missing ‘threshold’ in their framework—at what point should scenario studies be considered inconsistent. This thesis offers a re-interpretation on the concept of linking strategies by identifying the threshold for which scenarios can be considered inconsistent. In so doing, I would argue for the need to reinterpret Zurek and Henrichs (2007) concept of linking strategies to advance scholarship in multi-scale scenario research. This dissertation presents original research by developing an extension study on Canada’s energy futures under the Shared Socioeconomic Pathway (SSP) scenario framework. The SSP framework is intended to support more detailed analyses of societal change at a more localized scale; this framework is described in thematic special issues in Climatic Change and Global Environmental Change in 2014 and 2017 respectively. The SSPs described in these special issues are the ‘basic’ global version; from them, ‘extended’ SSPs could be elaborated further for detailed regional and national analyses (O’Neill et al., 2017, 2014). The basic SSPs provide a global framing for different socioeconomic and climate change policy developments up to 2100 (O’Neill et al., 2014). The Canadian oil and gas sector interacts directly with global energy markets and is already playing a key role in driving climate change, both as a high carbon emitter and as a major exporter of fossil fuels. Given this context, a multi-scale study provides an understanding of the broader implications of global influences on Canada’s low-carbon energy transition and vice-versa. According to the requirement set out in the SSP guidance note (van Ruijven et al., 2014), extension studies must be linked (or ‘hooked’) to the global SSPs in order to be consistent. The scientific community has developed multiple approaches for extending basic SSPs. One of the approaches is to re-specify the SSP elements. This extension study links to the SSP elements by adding elements necessary for more detailed national and sectoral analyses. Prior to developing scenarios for Canada, there is a need to identify relevant scenario elements. Identifying and prioritizing scenario elements are usually left to scenario developers’ subjective interpretation of experts or stakeholder opinions. How one expresses which scenario elements are important resides in individuals’ mental models, which are not accessible to others. In contrast, here candidate scenario elements are gleaned from the existing Canadian energy futures studies published in 2015 to 2016, which are then subjected to a network analysis. Network statistics can be used to more objectively identify which scenario elements are key since the method is transparent and data is accessible for public inspection (Lloyd and Schweizer, 2014). Elements identified as important by network analysis are then incorporated for multi-scale scenario analysis. Cross-impact balance (CIB) analysis (Weimer-Jehle, 2006) is used to search for scenario configurations that are consistent across scales. The result of multi-scale scenario analysis suggests that pathways to decarbonization in Canada are likely promoted by domestic effort regardless of which global development pathways (either carbonized or decarbonized) unfold. Scenarios in which the world remains carbonized and Canada decarbonizes and vice-versa are internally consistent. In relation to Zurek and Henrichs’ (2007) linking strategies, a conventional belief or assumption that global and local scenario outcomes must match across scales to be “consistent” has emerged in the scenario research community—though not everyone agrees with this assumption (e.g., van Ruijven et al., 2014; Wiek et al., 2013). This assumption was tested in this research. The result also tells us that internal consistency does not require that the outcomes across scales should be the same. Due to confusion about what cross-scale consistency means, there is the need to perform internal consistency checks in multi-scale scenario analysis. There is also the need to revise the operational definition of consistency across scales. The term scenario consistency across scales should not be confused with their degree of linkages (i.e., more or fewer links). Instead, we can use the consistency definition provided by CIB: internally logically consistent. Nonetheless, what may be more useful is to define the term “inconsistent”. This should be reserved for scenarios that are found to have internal logic problems—scenarios that, for good reasons, would be dismissed as implausible.